A General Class of Score-Driven Smoothers

Giacomo Bormetti, Department of Mathematics, University of Bologna

  • Date: 12 JULY 2018  from 14:30 to 16:00

  • Event location: Aula Seminari, 1st floor, Department of Statistica, Via Belle Arti 41, Bologna

  • Type: Statistics Seminars

Abstract: 

By interpreting score-driven models of Creal et al. (2013) and Harvey (2013) as approximate filters, we introduce a new class of simple approximate smoothers for nonlinear non-Gaussian state-space models that are named "Score-Driven Smoothers" (SDS). The newly proposed SDS improves on standard score-driven filtered estimates, as it employs all available observations. In contrast to complex simulations-based methods, the SDS has similar structure to Kalman backward smoothing recursions but uses the score of the non-Gaussian density. Through an extensive Monte Carlo study, we provide evidence that the performance of the approximation is very close to that of simulation-based techniques, while at the same time requiring significantly lower computational burden. 

Joint work with Giuseppe Buccheri, Fulvio Corsi, and Fabrizio Lillo

 

 L’Organizzatore                                                                                                         Il Direttore

Prof. Monia Lupparelli                                                                                        Prof. Angela Montanari   

 

 

La S.V. è invitata